1. Field of the Invention
The present invention relates to a technique of generating an image of a motion image of an object present in a virtual three-dimensional space using prestored motion data.
2. Description of the Related Art
In a three-dimensional video game, an image of an object, such as a character present in a virtual three-dimensional space, which is taking a motion, is generated by performing perspective transformation of the virtual three-dimensional space, including the object, with a virtual camera. To reproduce the motion of an object like a character with high fidelity, the object is made to take a motion using prestored motion data.
There are a variety of possible motions that may be executed by an object like a character. Preparing motion data for every possible motion results in a huge amount of motion data. In Japanese Patent Laid-Open Application No. 2003-62326 and Japanese Patent Laid-Open Application No. 2003-67773, motion data for movement and motion data for attacking are prepared as motion data of a character in a battle game. When a character attacks while moving, the motion data for movement and motion data for attacking are blended, and the character is caused to take a motion using the blended motion data. This can ensure reproduction of composite motions of a character with a smaller amount of data.
However, the schemes of blending motion data disclosed in the above-noted publications are limited to blending of motion data when different motions of a character, such as moving and attacking, are combined, and blending of motion data when one object and another object hit against each other. The schemes of blending motion data disclosed in the publications cannot reproduce a delicate difference between motions of an object which can be identified as essentially the same type. In addition, the schemes of blending motion data disclosed in the publications cannot freely change the motion of an object by the operation of a player.